Upload
karina-rosamelcacho
View
214
Download
0
Embed Size (px)
Citation preview
7/29/2019 B-2 Agonistas 3
1/14
Chapter 5 87
Chapter 5
Modeling of Conditions for the
Enantiomeric Separation of2-Adrenergic Sympathicomimetics byCapillary Electrophoresis using
Cyclodextrins as Chiral Selectors in a
Polyethylene Glycol Gel#
#
A slightly modified version of this chapter was published in:T. de Boer, R. Bijma and K. Ensing. J.Pharm.Biomed.Anal., 19 (1999) 529-537
7/29/2019 B-2 Agonistas 3
2/14
Chapter 588
Abstract
A two-factor central composite design was used to determine a
mathematical model for prediction of the optimal conditions for the
separation of the enantiomers of some widely used 2-sympathicomimetic
drugs (2-agonists) by capillary electrophoresis using cyclodextrins as a
chiral selector in a polyethylene glycol gel. The effects of the chemical
structure of these drugs along with the addition of polyethylene glycol to
the cyclodextrin solution on the resolution of their enantiomers were studied.
To allow impurity studies downto 0.1% (distomer/eutomer) a resolution of 2.5
should be warranted. Those 2-agonists containing two hydroxylic groups in
the aromatic ring structure show the highest enantiomeric separation, due
to the fact that one of their enantiomers has a better geometric structure to
fit into the -cyclodextrin cavity.
5.1 Introduction
2-Sympathicomimetics (2-agonists) stimulate the sympathic nervous
system by reducing the tone of smooth muscle cells particularly in the lungs
and the uterus. For that reason they are often used as vasodilator to
decrease arterial blood pressure (oxedrine), as bronchodilator in the
treatment of obstructive airway diseases (e.g. terbutaline, salbutamol) or as
tocolytic agent (i.e. fenoterol, ritodrine) to avoid premature birth.
Clenbuterol is a 2-agonist that is not authorized by the European Union as a
drug for human treatment but may be used for the treatment of bronchi-
obstruction in cattle. When the therapeutic doses of 2-agonists are
exceeded, these compounds act as growth promoters to improve meat-to-
fat ratios in cattle. Residues of these compounds, which are mostly
abundant in liver and in meat, are toxic to humans and can cause heart
complications. In addition, the medical commission of the International
Olympic Comity has listed the 2-agonists as illicit agents for athletes when
used in non-therapeutic doses.
The general molecular structure of these 2-agonists (terbutaline,
salbutamol, clenbuterol, fenoterol, ritodrine, oxedrine and isoprenaline)
7/29/2019 B-2 Agonistas 3
3/14
Chapter 5 89
along with their pKa-values is given in Table 5.1. These drugs are
administered as a racemate, but pharmacological studies have shown that
Table 5.1. Molecular structure of the used sympathicomimetics. The general
chiral center is depicted with *. A possible second chiral center
(ritodrine) is depicted with an #. The second chiral center in
fenoterol in the R1 substituent is depicted with *.
R4
R3
R2
N H
R1
R5
OH
*
#
pKa R1 R2 R3 R4 R5
Terbutaline
Salbutamol
Clenbuterol
Oxedrine
Isoprenaline
Ritodrine
Fenoterol
8.7
10.0
11.0
9.310.3
7/29/2019 B-2 Agonistas 3
4/14
Chapter 590
only one of the enantiomers has the desired therapeutic pharmacological
effect [1]. For that reason it is of great importance that the enantiomers of
such molecules can be fully baseline separated.
Basic enantiomers can be separated by capillary electrophoresis using
cyclodextrins (CD) as a chiral selector [2-8]. CDs are commercially available
oligosaccharides consisting of 6,7 or 8 D-(+)-glucopyranose units and they
are designated as , and -CD. Derivatization of the plain CDs leads to a
large variety of CDs, all with their own selectivity in their interaction with
chiral or non-chiral compounds. Derivatization of CD improves the solubility
in water/methanol solutions [9].
In chapter 4 we have shown that the enantiomers of terbutaline can
be readily separated resulting in a resolution that is satisfactory for
enantiomeric impurity studies (Rs > 2.5). In that study the addition of
polyethylene glycol 2000 (PEG-2000) showed a positive change in selec tivity
along with a decrease of the total amount of chiral selector necessary for
one separation. In this chapter it was tested if the observed phenomenon
would also occur for some terbutaline-analogs. For this reason, the optimal
conditions for the enantiomeric separation of the mentioned 2-agonists
were predicted by using a central composite design [10] as a
chemometrical tool to generate an empirical model. The use of
experimental designs in, for instance, pharmaceutical analysis has been
well described in reviews by Corstjens et al. [11] and Altria et al. [12]. More
specific was the use of a central composite design for the determination of
the optimum conditions for analysis of ranitidine [13], amphetamines [14]
and chlorophenols [15].
The hydroxypropyl--cyclodextrin (HP--CD) concentration and the PEG-
2000 concentration (two-variable design) were chosen as the parameters
for the selected design according to earlier described chiral optimization
models [5,16,17]. For this class of compounds HP--CD is the chiral selector
of choice (see Chapter 4).
7/29/2019 B-2 Agonistas 3
5/14
Chapter 5 91
5.2 Experimental
5.2.1 Apparatus
The CE system was a Model PRINCE with a 4 position sample tray and a
programmable injector system from Lauerlabs (Emmen, The Netherlands).
Detection at 210 nm was carried out with a LAMBDA 1000 UV/VIS VWL
detector (Bischoff, Leonberg, Germany). The fused-silica capillary with an
outer polyimide coating (50 m i.d., 375 m, o.d.) was from Polymicro
Technologies (Phoenix, AZ, USA). Data acquisition of CE/UV was performed
by the Maclab system (ADinstruments, Castle Hill, Australia) using the Chart
program (version 3.3, ADinstruments) for rec ording of the
electropherograms. For interpretation of the electropherograms, the Peaks
program (ADinstruments) was used. The vials used were 4 ml glass vials and
were obtained from PhaseSep (Waddinxveen, The Netherlands).
5.2.2 Chemicals and solutions
Ac etonitrile, methanol, NaOH and PEG-2000 all ofpro analyse (p.a.) quality
were obtained from Merck, Darmstadt, Germany. Hydroxypropyl--
cyclodextrin (HP--CD), p.a. was obtained from Wacker-Chemie, Munich,
Germany. The used 2-sympathicomimetics: terbutaline, salbutamol,
clenbuterol, fenoterol, ritodrine, oxedrine, isoprenaline, were all racemates
and of pharmacopoeial quality. Ritodrine and fenoterol were used in the
study as mixtures of four stereoisomers. They were dissolved in a solution
containing 1 part run-buffer and 9 parts of water to a final concentration of
20 g/ml. The CE run-buffer was prepared by dissolving sodium dihydrogen
phosphate monohydrate (p.a., Merck) to a concentration of 100 mM and
adjusting the pH with concentrated ortho-phosphoric acid (p.a. 85%,Merck)
to a pH of 2.5 giving a conductivity of 7.0 mS/cm. Water was purified with a
Milli-Q system (Millipore, Bedford, MA, USA). The conductivity of the purified
water was always less than 2 S/cm.
7/29/2019 B-2 Agonistas 3
6/14
Chapter 592
All solutions were filtered through a membrane filter (0.45 m) and
degassed for five minutes in an ultra-sonic bath (50 kHz, Branson Europa
B.V., Soest, The Netherlands), immediately prior to use.
5.2.3 CE conditions used for experiments
A capillary with a total length of 70 cm and an effective length of 55 cm
was used. An optical viewing window with a length of 0.5 cm, obtained by
burning off the polyimide coating, was aligned with the UV detection cell.
The coating of the first 2 mm of the capillary was also stripped.
New capillaries were rinsed with 1M sodium hydroxide for 10 minutes
at 1000mBar, with water for 10 minutes at 1000mBar and with the run-buffer
for 10 minutes at 1000mBar, respectively.
Cyclodextrins and polyethylene glycol were dissolved in the run buffer
and hydrodynamically injected as a removable gel until the capillary was
fully filled. The latter was monitored by UV-detec tion. The analytes then
were electrokinetically injected in order to avoid the PEG/CD gel to be
forced out. The injection (10 kV, 6s) and separation voltage (30 kV) were
ramped at 6 kV/s and took place at a constant temperature of 15C. After
each run the capillary was refilled with fresh PEG/CD gel. Because we inject
the removable liquid gel directly into the capillary instead of adding it to
the ground-electrolyte, the electro-osmotic flow (EOF) should be suppressed
to avoid the removable gel to be forced out of the capillary. The latter is
partially accomplished by the increased viscosity, but a working pH of the
ground-electrolyte of pH
7/29/2019 B-2 Agonistas 3
7/14
Chapter 5 93
started when the ground electrode and the capillary-end were placed into
the vial containing the run buffer.
5.3 Statistical methods for experiments
Chemometrical analysis was performed with a Design-Expert program,
version 3.05 (Stat-Ease, Inc, Minneapolis, MN, USA). Contour-plots were
produced by Matlab: high performance numeric computation and
visualisation software (Natick, MA, USA).
5.4 Results and discussion
A two-factor central composite design was used to obtain data for the
fitting of a second-order polynomial model that is defined as [10]:
y1i = b0 + b1x1i + b2x2i + b11x1i2 + b22x2i2 + b12x1ix2i +e1i
where x1 and x2 indicate the representative variables (quantitative factors):
percentage PEG-2000 and concentration HP--CD, respectively. A two-
factor design (i.e. x1 and x2 ) results in 2k + (2*k) + 1 = nine experimental
conditions. In order to estimate the experimental uncertainty, the center
point of the design is measured another three times, resulting in twelve
experiments for each sympathicomimetic drug that take place in random
order. The window wherein the design was selected was restricted by the
maximum concentrations of PEG-2000 (10%) and CD (25 mM). Larger
concentrations gave rise to strong baseline fluctuations with the
consequence that the efficiencies became very low or even that peaks
could not be detected. The corner points of the design were selected
according to previous experiments for the enantiomeric separation of
terbutaline (see Chapter 4).
7/29/2019 B-2 Agonistas 3
8/14
Chapter 594
For example, the experimental conditions for the enantiomeric separation
of clenbuterol were tested in random order and are presented along with
the corresponding resolutions (y1i) in Table 5.2. The coefficients of the
mathematical model were obtained by matrix calculation via b = (XX)-1
(Xy), where X is a matrix containing the columns which contain the
coefficients of the model parameters for each experiment. Finally the
polynomial can be given
Rs = -2.1 0.07x1i + 0.36x2i - 0.11x1i2
- 0.14x2i2
+ 0.28x1ix2i + e1i.
The adequac y of the model, the goodness of fit is summarized by the
coefficient of multiple determination (R2 = 0.846) and by the F-ratio for lack
of fit (FLOF = 0.05). Using a 95% significance level the model is probably
adequate, i.e. there is no lack of fit, when the FLOF 0.05 [10]. It is important
to realize that the polynomial is only valid for the tested ranges, i.e. 3-9%
PEG-2000 and 8 - 22 mM HP--CD. The results of the calculated polynomial
are visualised as a contour plot in Fig. 5.1. In the same manner experiments
Table 5.2 Experimental conditions for the enantiomeric separation of
clenbuterol along with the obtained resolutions and the calculatedpolynomial
PEG 2000 (%) = x1 HP--CD(mM) = x2 Rs
6
6
6
64
8
8
9
6
64
3
15
15
15
1510
10
20
15
8
2220
15
2.1
2.2
2.0
2.21.5
1.7
1.6
2.0
1.2
2.62.5
1.9
Rs = -2.1 0.07x1i + 0.36x2i - 0.11x1i2
- 0.14x2i2
+ 0.28x1ix2i + e1i
7/29/2019 B-2 Agonistas 3
9/14
Chapter 5 95
were completed for the other 2-agonists. The resulting polynomials, the
concentrations needed to warrant the aimed resolution of 2.5 or higher and
the adequacy of the models are given in Table 5.3
As can be seen in Table 5.3, the calculated models for isoprenaline and
oxedrine exhibit a significant lack of fit (FLOF = 0.016 and 0.015, respectively).
For a good separation of the enantiomers of the 2-agonists, a three point
interaction or more of the hydroxyl and amino groups of one of the
sympathicomimetic enantiomers with the hydroxyl groups of the
cyclodextrins and a complexation of the (substituted) aromatic ring of this
enantiomer with the inner hydrophobic moiety of the cyclodextrin cavity is
necessary [22,23]. We try to explain the data given in Table 5.3 combined
with the molecular structures and their pKa values given in Table 5.1.
Because the pKa values of the tested compounds are practically the same
and all compounds are fully protonated at the working pH (2.5), differences
Table 5.3. The calculated polynomials with their adequacy and predictive ability of
the -agonists along with the conditions that will warant a resolution of 2.5or highera.
Cat Compound Results of the second order
polynomial
PEG 2000
(%) = x1
HP--CD
(mM) =x2
Rs
A Terbutaline R2 = 0.884; LOF = 0.153; 10.0* 25* 7.9
A Isoprenaline R2 = 0.869; LOF = 0.016; 10.0* 25* 2.7
A Fenoterol R2
= 0.824; LOF = 0.068; 4.7 11.7 2.0
B Salbutamol No separation
C Ritodrine R2
= 0.592; LOF = 0.389; 6.0 25* 1.5
C Oxedrine R2 = 0.838; LOF = 0.015; 10.0* 15.5 1.1
D Clenbuterol R2 = 0.846; LOF = 0.050; 0.0* 25.0* 4.2
a Polynomials are valid for the ranges 3-9% PEG 2000 and 8-22 mM HP--CD.* When the predicted concentrations are outside the tested ranges, the values are marked
with an asterisk
7/29/2019 B-2 Agonistas 3
10/14
Chapter 596
in resolutions can be attributed to differences in their molecular structure
only.
We divided the analyzed compounds in four categories. A) compounds
with two phenolic-hydroxy groups in the aromatic ring structure. B)
compounds with a phenolic hydroxy group and an aliphatic hydroxy group
in the aromatic ringstructure. C) compounds with one phenolic hydroxy
group in the aromatic ringstructure. D) other substituents in the aromatic
ringstructure.
To allow enantiomeric impurity profiling down to 0.1% a resolution of
2.5 is necessary, as can be calculated from the theoretical overlap of
symmetrical peaks by simple overlap studies as mentioned in the former
Chapter. The resolution is defined as the extent of separation between two
compounds and is formally calculated by R = 2 (t2 - t1 ) / (w1 + w2 ) [24],
where t 1 and t 2 are the migration times and w 1 and w 2 the peakwidths at
the baseline of the more mobile (1) and the less mobile (2) analyte.
According to Table 5.3 compounds belonging to category A (fenoterol,
terbutaline and isoprenaline) and to category D (clenbuterol) give the best
resolution. It appears that the largest discrimination between the
enantiomers is found were the 2-mimetics contain substituents in the
aromatic ringstructure and espec ially on the R2 and R4 position (terbutaline
and c lenbuterol), due to the fac t that one of their enantiomers has a better
geometric structure to fit into the -cyclodextrin cavity.
In Figures 5.1(I) and 5.1(II) it is shown that for clenbuterol and
terbutaline, respectively, the minimal resolution we were aiming for (2.5) is
readily achieved within the tested design without the necessity for further
optimization.
In Fig 5.1(II) it can also be seen that the addition of PEG-2000, referring to
the contour responding to a resolution of 2.5, results in a decrease of the
necessary concentration of CDs. The same contour in Fig. 5.1(I) shows a
different effect of the addition of PEG 2000. This phenomenon can not be
explained at the moment. In both cases, however, the optimal conditions
for a satisfactory result (i.e. a resolution of 2.5) should be achieved with the
lowest concentration possible for either component.
7/29/2019 B-2 Agonistas 3
11/14
Chapter 5 97
For isoprenaline, which contains two lipophilic substituents located on R2
and R3, the minimal resolution cannot be obtained within the selected
design, but is still situated within the restricted window area (10% PEG-2000
and 25 mM CD), therefore further optimization is necessary. The latter could,
for instance, be accomplished by using a two-factor simplex experimental
design [10,11]. Fenoterol which also has the apparently essential lipophilic
substituents on R2 and R4 also contains a large R1 substituent that appears
to decrease the resolution compared to the above described compounds.
Although fenoterol has two chiral centres, only one pair of enantiomers can
Fig.5.1. The calculated contourplots of the -agonist belonging to category A and D.(I) terbutaline; (II) clenbuterol; (III) fenoterol; (IV) isoprenaline. The central composite
design used is designated in the figure.
7/29/2019 B-2 Agonistas 3
12/14
Chapter 598
be separated within the used maximum separation time (i.e. 30 minutes). A
resolution of 2.5 in this case cannot be accomplished within the used
restricted window area.
Apparently, the methoxygroup of salbutamol on the R2 position, category
B, prevents separation of the enantiomers because no resolution is
obtained. When we compare the compounds of category C with the
compounds of the other categories, we observe that for both compounds
no baseline separation can be obtained.
5.5 Conclusion
The use of HP--CD as a chiral selec tor in a polyethylene glycol solution
appears to be an adequate method to separate the enantiomers of the 2-
sympathicomimetics. The addition of PEG changes the selectivity but does
not always result in an increase in resolution. In most cases addition of PEG
results in a lower concentration of CDs necessary to obtain the desired
minimum resolution of 2.5 necessary for the detection of impurities as low as
0.1 % (distomer / eutomer).
The enantiomers of compounds containing two hydroxylic groups
substituted at the aromatic ring could be readily separated, resulting in the
aimed resolution, due to a higher complexation with the cyclodextrins of
one of the enantiomers. The selected design within the restricted window
area (10% PEG-2000 and 25 mM CD) appears to be adequate for
calculation of a mathematical model for a fast optimization of the
separation. If the desired resolution is not achieved within the valid range of
this two-factor composite design, further optimization could be carried out
with a two-factor simplex design. A better resolution for the compounds of
category B and C can probably be achieved by using other types of
cyclodextrins.
5.6 Acknowledgements
The authors would like to thank Dr. P.M.J . Coenegracht for valuable
discussions.
7/29/2019 B-2 Agonistas 3
13/14
Chapter 5 99
5.7 References
1 A.B. J eppson, K. J ohansson and B. Waldec k. (1984) Steric aspects of agonism and
antagonism at beta-adrenoreceptors: experiments with the enantiomers of terbutaline,
Acta Pharmacol.Toxicol.54, 285-2912 Th.L. Bereuter.(1994)Enantioseparation by capillary electrophoresis, LC-GC Int.2, 78-93
3 S. Fanali. (1991) Use of cyclodextrins in capillary electrophoresis. Resolution of
terbutaline and propranolol enantiomers,J .Chromatogr.545, 437-4444 S. Fanali and E. Camera. (1996) Use of methylamino--cyclodextrin in capillary
elec trophoresis. Resolution of acidic and basic enantiomers, Chromatographia.43, 247-
253
5 A. Guttman and N. Cooke. (1994) Practical aspects of chiral separations of
pharmaceuticals by capillary electrophoresis. I. separation optimization,J .Chromatogr.A.680, 157-162
6 A. Guttman, S. Brunet and N. Cooke. (1996) Capillary elec trophoresis separation of
enantiomers using cyclodextrin array chiral analysis, LC-GC Int.2, 88-1007 I.S. Lurie, R.F.X. Klein, G.A. Dal Cason, M.J . LeBelle, R. Brenneisen and R.E. Weinberger.
(1994) Chiral resolution of cationic drugs of forensic interest by capillary electrophoresis
with mixtures of neutral and anionic cyclodextrins, Anal.Chem.66, 4019-4026
8 L.A. St.Pierre and K.B. Sentell. (1994) Cyclodextrins as enantioselective mobile phase
modifiers for chiral capillary electrophoresis. Effects of pH and cyclodextrinconcentration,J .Chromatogr.B.657, 291-300
9 R. Kuhn, F. Stoecklin and F. Erni. (1992)Chiral separations by host-guest complexation
with cyclodextrin and crown ether in capillary zone electrophoresis, Chromatographia.33, 32-36
10 D.L. Massart, B.G.M. Vandeginste, S.N. Deming and L. Kaufman, Chemometrics: a
Textbook. Datahandling in Science and Technology 2, Elsevier, Amsterdam, The
Netherlands, 1988.11 H. Corstjens, H.A.H. Billiet, J . Frank and K.C.A.M. Luyben.(1995)Optimisation of selectivity
in capillary electrophoresis with emphasis on micellar electrokinetic capillary
chromatography,J .Chromatogr.A.715, 1-11
12 K.D. Altria, B.J . Clark, S.D. Filbey, M.A. Kelly and D.R. Rudd. (1995) Application ofchemometric experimental designs in capillary electrophoresis: a review,
Electrophoresis.16, 2143-2148
13 V.M. Morris, C. Hargreaves, K. Overall, Ph.J . Marriott and J .G. Hughes. (1997)
Optimization of the capillary electrophoresis separation of ranitidine and relatedcompounds,J .Chromatogr.A.66, 245-254
14 E. Varesio, J .Y. Gauvrit, R. Longeray, P. Lanteri and J .L. Veuthey. (1997) Central
composite design in the chiral analysis of amphetamines by capillary electrophoresis,
Electrophoresis.18, 931-93715 M. J imidar, P.F. De Aguiar, S. Pintelon and D.L. Massart.(1997)Selectivity optimization for
the separation of chlorophenols in an irregularly shaped experimental region in capillary
electrophoresis,J .Pharm.Biomed.Anal.15, 709-72816 Y.Y. Rawjee and G. Vigh. (1994) A peak resolution model for the capillary
electrophoretic separation of the enantiomers of weak acids with hydroxypropyl--
cyclodextrin-containing background electrolytes, Anal.Chem.66, 619-627
17 S.A.C. Wren and R.C. Rowe.(1992)Theoretical aspects of chiral separation in capillaryelectrophoresis. I. Initial evaluation of a model,J .Chromatogr.603, 235-241
18 M. Albin, P.D. Grossman and S.E. Moring. (1993)Sensitivity enhancement for capillary
electrophoresis, Anal.Chem.65, 489A-497A
19 R.L. Chien and D.S. Burgi.(1991)Field amplified sample injection in high performancecapillary electrophoresis,J .Chromatogr.559, 141-152
7/29/2019 B-2 Agonistas 3
14/14
Chapter 5100
20 A. Guttman and H.E. Schwartz. (1995) Artifacts related to sample introduc tion in
capillary gel electrophoresis affecting separation performance and quantitation,
Anal.Chem.67, 2279-2283
21 X. Huang, M.J . Gordon and R.N. Zare. (1988) Bias in quantitative capillary zoneelec trophoresis caused by electrokinetic sample injec tion, Anal.Chem.60, 375-377
24 C.E. Dalgliesh. (1952) The optical resolution of aromatic amino-ac ids on paper
chromatograms,J .Chem.Soc .137, 3940-3942
25 A. Guttman, A. Paulus, A.S. Cohen, N. Grinberg and B.L. Karger. (1988) Use ofcomplexing agents for selective separation in high-performance capillary
electrophoresis. Chiral resolution via c yclodextrins incorporated within polyacrylamide
gel columns,J .Chromatogr.448, 41-53
26 United States Pharmacopoeial Convention, United States Pharmacopoeia, XXIII edition,9th supplement, Rockville (MD), USA, 1998.